EuroEval Spanish NLU - ScaLA ES: leaderboard
The "ScaLA ES" subtask of EuroEval Spanish NLU. Spanish-language EuroEval natural-language-understanding suite, separating NLU task performance from the broader all-task EuroEval aggregate.
Metric: Matthews correlation coefficient (%). Source: euroeval.com. Saturation forecast: Around September 2028. 413 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 3 Pro (Preview) | 73.71 | #66 |
| 2 | Claude Sonnet 4.6 | 69.02 | #90 |
| 3 | Gemini 3 Flash (Preview) (Non-reasoning) | 68.43 | #82 (Gemini 3 Flash (Preview)) |
| 4 | Gemini 3 Flash (Preview) (Thinking) | 62.92 | #82 (Gemini 3 Flash (Preview)) |
| 5 | GPT-5.4 Mini (2026-03-17) (High) | 62.38 | #169 (GPT-5.4 Mini (2026-03-17)) |
| 6 | GPT-5.2 | 61.92 | #110 |
| 7 | Gemini 3.1 Flash Lite (Preview) | 61.41 | #194 |
| 8 | Qwen 3.5 35B A3B Base | 57.97 | #310 |
| 9 | Llama 3.1 405B | 56.52 | #466 |
| 10 | GPT-5.4 Nano (2026-03-17) (High) | 55.08 | #266 (GPT-5.4 Nano (2026-03-17)) |
| 11 | Qwen 3.5 9B | 52.26 | #379 |
| 12 | Grok 4.1 Fast (Reasoning) | 52.16 | #217 (Grok 4.1 Fast) |
| 13 | Qwen 3 4B 2507 (Thinking) | 49.21 | #538 (Qwen 3 4B 2507) |
| 14 | NVIDIA-Nemotron-3-Nano-30B-A3B-BF16 | 48.02 | #679 |
| 15 | Qwen 3 235B A22B 2507 FP8 (Thinking) | 47.39 | #246 (Qwen 3 235B A22B 2507 FP8) |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=euroeval-spanish-nlu-scala-es · How It Works · Data refreshed daily, snapshot 2026-10-20.